The models are becoming the commodity → brilliant, abundant, increasingly interchangeable. The data they train on and act through is not. Creation, storage, deployment, security, orchestration: that pipeline is where the switching costs live, and switching costs are the only moat software ever really had. $RDDT $INOD $NOW $SNOW $CRWD $DDOG $PLTR Source: RosannaInvests on […]
The AI data layer is the moat Read More »
Bank of America is calling for three rate hikes by year-end. The data says that call is wrong. Core CPI is up just 0.2% month over month. Core PPI is below consensus. One-year breakevens are 1.698%, down 25.78% year to date. The oil curve is moving into contango, which says supply shock rather than persistent
Bank of America Is Wrong About Three Rate Hikes Read More »
Inference is the next phase of AI, and inference is bottlenecked by memory, not compute. Training gets the headlines because it demands giant clusters. Inference is different: it is the recurring work of serving models to real users, agents and applications. Every request needs fast memory close to the GPU, and every delay in moving
Inference is bottlenecked by memory, not compute Read More »
AI’s bottleneck is no longer chips. It is power, and the grid is tapped out. Gas turbines: backordered to 2030. Nuclear: a decade out. Solar plus storage: built in months. $TE builds both on US soil, and demand for its 2027 to 2028 output already exceeds everything it can make.
AI’s Bottleneck Is Power, Not Chips Read More »
12 companies built to grow revenue tremendously over the next three years. Not the mega-caps everyone owns. The layers underneath them, across compute, photonics, energy, physical AI and space. Growth rates below. Stay to the end for the part that actually matters: what you pay for it.
Built to grow: 12 companies and FY2026e revenue growth Read More »
Growth can decelerate without destroying the investment case. A delayed milestone or a back-half-weighted year matters, but it is not automatically a broken thesis. The key distinction is between timing and economics. Investors should separate the path from the destination. A slower ramp changes present value, financing needs, and the amount of patience required. It
Growth Can Decelerate Without Breaking the Thesis Read More »
$DGXX is following the exact $APLD playbookabout 18 months behind. The pattern institutional investors are not pricing yet is the identical progression: both companies started as crypto miners, pivoted into AI data centers, and signed transformative anchor tenants. That matters because the pivot is not a story change; it is a monetization change. Once a
DGXX Is Following the APLD Playbook: 18 Months Behind Read More »
🚨This is the most bullish thing that's happened to AI infrastructure all year – it’s financial validation! Google isn't dumping shares. It's doing a textbook raise-to-GROW. There's a difference between raising equity to SURVIVE and raising to fund demand you physically can't meet. $GOOGL is the second kind and the tell is Berkshire. Buffett's shop
The most bullish thing for AI infrastructure Read More »
Outside post-recessionary rebounds -> but we did have a recession! It just rolled through manufacturing (26 straight months of ISM contraction <50 – the longest on record), freight and goods rather than hitting all at once. Cap-weighted EPS hid it. The rebound fits the S&P 500 earnings are now expected to increase by 24% this
Outside post-recessionary rebounds: but we did have a recession! Read More »
Credo Technology just closed a year most companies will never seeand the market response was red. That disconnect is the thesis: investors are treating extraordinary growth as if it is already decelerating. Revenue more than tripled to $1.3 billion for FY26, while non-GAAP net income increased more than fivefold to $662 million. Q4 revenue reached
Credo’s Triple-Digit Growth Is Being Priced Like a Slowdown Read More »
